2 1 8  Fluid  Extracts  by  Reper eolation.     {  %^y%87h8arm' 
every  successive  percolation  tends  to  correct  the  errors  and  improve  the 
accuracy  of  those  which  have  preceded  it. 
For  the  purposes  of  this  paper  as  a  model,  and  to  construct  the 
following  table,  the  percolate  from  the  first  five  percolations  of  this 
series  was  divided  into  much  smaller  fractions  than  is  directed  in  the 
process,  or  needed  in  actual  practice,  leaving  for  a  subsequent  table 
the  model  for  actual  practice.  The  fractions  of  the  first  5  were 
somewhere  near  2  parts  or  2  ounces  each, =56*7  grams.  It  was 
not  easy  to  get  them  uniform  in  weight,  and  therefore  this  was  not 
attempted.  Each  fraction  was  shaken  up  and  accurately  weighed. 
Then  about  1  cubic  centimetre  was  measured  off  into  a  tared 
shallow  watch  glass,  and  quickly  weighed  to  avoid  loss  by  evaporation. 
These  weighed  portions  were  then  dried  by  steam  heat  until  they 
ceased  to  lose  weight  and  the  residue  carefully  weighed.  This  gave 
the  data  for  calculating  the  total  residue  which  each  fraction  would 
give  if  the  whole  fraction  had  been  evaporated.  But  a  large  pro- 
portion of  this  residue  was  glycerin,  and  it  was  sought  to  get  at  the 
extract  only.  So  a  cubic  centimetre  of  the  menstruum  was  weighed 
and  evaporated  in  the  same  watch  glass  until  it  ceased  to  lose  weight 
under  the  same  conditions,  and  it  was  found  that  the  residue  weighed 
16  per  cent,  of  the  weight  of  the  menstruum.  This  was  an  unex- 
pected result,  for  as  the  menstruum  contained  20  per  cent,  of  its  weight 
of  glycerin,  it  was  to  be  expected  that  upon  evaporation  very  nearly 
this  proportion  would  remain  on  the  watch  glass.  Therefore  the  trial 
was  three  times  repeated,  and  the  mean  of  the  trials  was  within. a  very 
small  fraction  of  16  per  cent.,  showing  that  about  4  per  cent,  of  the 
glycerin  went  off  with  the  alcohol  and  water.  This  gave  the  remain- 
ing element  for  closely  approximating  the  extract  in  each  fraction  by 
the  following  formulas.  As  the  weight  of  the  cubic  centimetre  taken- 
from  each  fraction  of  percolate  is  to  the  weight  of  the  residue  after 
evaporation,  so  is  the  weight  of  the  entire  fraction  to  the  weight  of  the 
residue  it  would  leave  on  evaporation.  But  had  this  fraction  been- 
menstruum  instead  of  percolate,  and  been  evaporated  under  the  same 
conditions,  it  would  have  given  16  per  cent,  of  its  weight  of  glycerin 
as  a  residue.  It  was  therefore  assumed  (though  of  course  not  cor- 
rectly) that  16  per  cent,  of  the  entire  fraction  was  glycerin.  Then  16 
per  cent,  of  the  weight  of  each  fraction  was  taken  and  this  amount 
